Guest blogger, Alyson Cartwright, Product Manager
Numerical data analysis is an important step in understanding geologic characteristics of mineral deposits and/or building geologic models. It permits geologists to study characteristics of the deposit based on the sampled data, as well as to validate the consistency of the different models once they are built. In mining, based on the methods implemented, numerical data analysis is typically classified into two categories: 1) statistical and 2) geostatistical analysis. The statistical analysis focuses on studying the distributions of metal content attributes and the relationship between them in different geologic scenarios. This analysis is important to be able to control the reproduction of features identified in the sampled data in the models of the deposit. The geostatistical analysis focuses on studying the spatial continuity of metal attributes in different geologic scenarios. This analysis is important to design proper estimation plans for calculating metal content at unsampled locations in the models of the deposit.
Sigma provides tools to assist geologists to carry out numerical data analysis of geologic information. Each of these tools generates a graph and a report of the corresponding summary statistics. The assist functionality comes in the form of algorithms that identify relevant features of the data source. In each graph, these relevant features are used to calculate default parameters. The assist functionality significantly reduces the time spent analyzing the geologic information as most of the graph calculation parameters have been pre-computed and presented as default settings. For example, histograms can be calculated by only indicating the item name in the data source. Sigma suggests the bin definition and frequency limits. Similarly, directional experimental variograms can be calculated by simply indicating the item name in the data source along with azimuth and dip angles to indicate the direction. Sigma suggests the lag configuration and corresponding tolerances.
Figure 1: Screenshot of the interface.
For exploratory data analysis purposes, Sigma provides a quick graph functionality that allows calculating sets of histograms and cumulative probability plots, for different combinations of categorical information in a single run. This functionality saves time as it is not necessary to generate all these graphs individually. The categorical information can be set as a list, for one categorical attribute, or as a matrix, for two categorical attributes. This analysis permits you to obtain an initial snapshot of how the different geologic categories act as controls of the different metal grade items in the deposit.
Figure 2: Sigma graphs: A) variogram map, B) histogram, C) quantile-quantile plot, and D) scatter-plot frequency map.
The supported data sources are: drillhole, blasthole, and point samples (MineSight Torque and legacy files), block models (regular blocks, sub-blocks, and gridded seams), and data stored in ASCII files. The statistical tools in Sigma are: histograms, cumulative probability plots, grade-tonnage curves, box-plots, scatter-plots, quantile-quantile plots, and swath-plots. These tools are used to study characteristics of univariate distributions, compare univariate distributions of one or two different data sources, study characteristics of bivariate distributions, and to compare the summary statistics of block models in different regions with specific orientations. The geostatistical tools in Sigma are: variogram maps, directional variograms, global variograms, downhole variograms, and variogram fitting. These tools are used to fit variogram models based on an analysis of the three-dimensional variogram space and one-dimensional directional projections.
Sigma can interact with MineSight 3D to identify samples that correspond to specific regions in distributions. This functionality allows visualizing in three dimensions the position of the samples that fall within regions of a distribution, univariate, or bivariate. This three-dimensional visualization permits you to identify the presence of trends in the data source. For example, verify if low grade samples are preferentially located in certain regions of the domain or evenly distributed.
Figure 3: Samples identified to be within a histogram bin interval in Sigma are highlighted in yellow in MS3D.
Sigma makes it significantly easier to carry out numerical data analysis of geologic information by allowing geologists to directly focus on studying different data sources. Sigma reduces the complexity of using the different tools by suggesting calculation parameters, which are estimated by algorithms that identify principal characteristics of the data sources. The batch functionality allows processing a vast range of geologic scenarios quickly. The interaction with MS3D makes the analysis dynamic, thus, allowing you to quickly make a decision or take corrective measures. These features make Sigma an efficient product for analyzing geologic information.